N. G. Esipova

On the optimal folding of protein molecules

The process of formation of a globular structure by a long molecular chain has been examined. In this process, various regions of the chain interact with one another. We classify the contacts thus formed as “correct” and “erroneous” ones. The correct contacts are those characteristic of the final native globular structure. All other contacts can be treated as erroneous. It is demonstrated that globule formation may proceed actually without formation and subsequent decay of erroneous contacts. Our model permits avoiding examination of numerous erroneous variants inasmuch as the regions of the chain that form correct contacts enter “long-range” interactions that at the same time can be highly selective. The existence of interactions of this kind facilitates the mutual approach and interaction of just those regions of the chain that yield correct contacts. Based on database analysis, it is shown that the model is valid not only for abstract structures but also for real polypeptide chains capable of forming protein globules and helical fibrils.

Abundance and arrangement of single, double and bifurcated hydrogen bonds in protein α-helices: Statistical analysis of PDB-select

Statistics are collected and analyzed for the possibility of hydrogen bonding in the secondary structures of globular proteins, based on geometric criteria. Double and bifurcated bonds are considered as pairs of admissible H-bonds with two proton donors or two proton acceptors, respectively. Most of such bonds belong to peptide groups in α-helices, with Oi…Ni + 3 nearly as frequent as Oi…Ni + 4; in contrast, most of the 3/10-helical segments are too short to have any. Alternating double and bifurcated bonds in α-helices form an apparently cooperative network structure. A typical α-helical segment perhaps carries two stretches of the H-bond network broken in the middle. The constituent H-bonds are nonlinear: the hydrogen atom is off the straight line connecting the proton donor and proton acceptor atoms. This deflection is larger for Hi + 3 vs. bond line Oi−Ni + 3 than for Hi + 4 vs. Oi−Ni + 4, and though the two kinds of bond have about the same length (exceeding those typical of low-molecular compounds), Oi…Ni + 4 must be stronger than Oi…Ni + 3. Double/bifurcated bonds are also not coplanar, i.e., hydrogen atoms are beyond the N…O…N (or O…N…O) plane.

Mutual disposition of short conformationally stanch oligopeptides in the 3D structure of globular proteins

The disposition of conformationally stanch, helix-prone tetrapeptides and of longer segments containing them in proteins of different structural and functional groups (PDBselect and CATH samples) has been analyzed. Quasirandom Monte Carlo experiments show that the disposition of such segments can be regarded as stochastic. At that, ∼60% of stanch peptides in the protein globules have at least one stanch neighbor within 5 Å.

Alternatingly twisted β-hairpins and nonglycine residues in the disallowed II′ region of the Ramachandran plot

The structure of the SH3 domain of α-spectrin (PDB code 1SHG) features Asn47 in the II′ area of the Ramachandran plot, which as a rule admits only glycine residues, and this phenomenon still awaits its explanation. Here, we undertook a computational study of this particular case by means of molecular dynamics and bioinformatics approaches. We found that the region of the SH3 domain in the vicinity of Asn47 remains relatively stable during denaturing molecular dynamics simulations of the entire domain and of its parts. This increased stability may be connected with the dynamic hydrogen bonding that is susceptible to targeted in silico mutations of Arg49. Bioinformatics analysis indicated that Asn47 is in the β-turn of a distinctive structural fragment we called ‘alternatingly twisted β-hairpin.’ Fragments of similar conformation are quite abundant in a nonredundant set of PDB chains and are distinguished from ordinary β-hairpins by some surplus of glycine in their β-turns, lack of certain interpeptide hydrogen bonds, and an increased chirality index. Thus, the disallowed conformation of residues other than glycine is realized in the β-turns of alternatingly twisted β-hairpins.

The structural and physicochemical characteristics of conformationally stable α-helical oligopeptides

An analysis of conformationally stable (conformationally conservative) tetrapeptides selected from protein structures that are deposited in the PDBSelect database has been carried out. The set contained 943 different tetrapeptide sequences, with each sequence occurring at least five times in different fragments of protein structure. Analysis of conformations based on DSSP markup revealed that the conformation of most peptides (900 of 943 sequences) was α-helical, while 43 sequences had different conformations, of which the left-handed polyproline II helix is especially worth mentioning. The physicochemical properties of conformationally stable peptides from each subset were inferred from the average hydropathicity of the tetrapeptides. The results of the calculations revealed “neutral” hydropathicity of the conformationally stable oligopeptides. Notably, the distribution of hydropathicity values for the conformationally stable peptides was considerably narrower than those for the control sets of peptides. Thus, conformationally stable oligopeptides form a specific group of local protein structures with nearly uniform conformational and physicochemical properties. The theory of specific long-range interactions that was previously developed by the authors of the present study assumes such peptides to be well adapted for efficient mutual recognition of molecules.

Long-distance interactions and principles of molecular recognition at various biosystem organization levels

Efficient molecular recognition, in which recognition processes are occurring much faster than it takes to test variants, is only possible when long-distance recognition occurs together with contact interactions. The distance between interacting molecules should be sufficiently long to prevent hindrances to the search and, on the other hand, sufficiently short to provide selectivity. It was demonstrated that both of these two requirements can be satisfied simultaneously for biological macromolecules that include helical segments. Because the “diameters” of helical molecules are far shorter than their lengths, the intermolecular distance can be far greater than the diameters, thus allowing a free search. The distance can be far shorter than the lengths at the same time, thus providing selectivity. Analytical procedures were developed to estimate the parameters for protein–protein and protein–nucleic acid recognition. The coincidence of the charge-distribution periods in helical segments was found to substantially increase their interaction potential, and the reduction scale characteristic of the potential was shown to depend on the numerical value of the coinciding period.

Conserved double-stranded DNA regions (“cophased blocks”) of transcriptional regulatory modules are close in space due to phasing relative to the nucleosome DNA superhelix

Interspecies alignment of homologous nucleotide sequences at gene loci controlling the early development of Drosophila has resulted in identification of conserved domains that range in length from 30 to 70 nucleotides and are, in this respect, intermediate between transcription factor binding sites (usually about 7–10 nt) and nucleosome repeat units (165–210 nt). These domains are located mainly in the region of known functional elements (cis-regulatory modules) of the locus (enhancers, the proximal promoter, and the coding segment). Taken together, they occupy no more than half of total enhancer length but contain the absolute majority of annotated binding sites for transcription factors. Their distribution has a quasi-periodic pattern that agrees well with the pattern of experimental localization of nucleosomes. The distance between neighboring domains is about 84 nt, which is equivalent to the pitch of the nucleosomal DNA superhelix. This regularity allows the identified conserved domains to be regarded as “cophased blocks” that are spatially converged due to location in the neighboring coils of the nucleosome DNA superhelix.

Disallowed conformations of a polypeptide chain as exemplified by the β-turn of the β-hairpin in the α-spectrin SH3 domain

A comprehensive conformational analysis has been performed of β-turns containing an amino acid with a disallowed conformation of the polypeptide chain backbone. The first residue of the β-turn (Asn47) of the distal β-hairpin in the α-spectrin SH3 domain is characterized by a sterically disallowed back-bone conformation, with values of dihedral angles Φ and ψ being in the right bottom quadrant of the Ramachandran plot). Based on analysis of all α-spectrin structures with an anomalous element deposited in the PDB, the hypothesis has been proposed that the disallowed conformation may result from the fixation (conditioned by the SH3 domain structure) of β-structure residues adjacent to the β-turn, which are arranged so that such a residue can adopt only a disallowed conformation, excluding the possibility of other β-turn conformations (with an allowed local conformation). To test this hypothesis, a conformational analysis of the β-turn was performed by varying all its internal coordinates (two pairs of Φ and ψ angles and two ω angles). As a result of a grid search procedure with a 1° step, variants were selected that corresponded to stereochemically allowed local deformations in the polypeptide chain segment forming the β-turn. However, the local conformation of the Asn47 residue in all these variants proved to remain in the disallowed region. These variants included conformations coinciding with experimentally determined strictures from the PDB as well as an additional variant differing from them in the pair of Φ and ψ angles at the second residue of the β-turn: in the Ramachandran plot, they fall into the region near the line Φ = 0 and negative ψ values, i.e., into the strongly disallowed region where no experimental points have been recorded. Thus, the results of this study confirm the hypothesis that the disallowed conformation is “imposed” on the β-turn due to fixation of residues adjacent to it. Topological structural limitations in β-hairpins of such a type exclude the possibility of allowed local conformations.

Conformationally stable segments in helical structures of polypeptide chains of proteins and their role in high level structures formation

In the work the arguments are presented in favor of the idea on the role of conformationally stable oligopeptides in specific long-distance interactions in phenomena of molecular recognition during various biological processes. Original authors’ and literature data are taken into account. The examples of conformationally stable short oligopeptides participating in alpha-helix and collagen type structures formation are given simultaneously with theoretical approaches. The conformationally stable oligopeptides obtained in the course of PDB bank analysis are discussed. The role of amino acid sequence in collagen helix formation is shown.

Left helix of polyproline II type and genesis of β-structures in spidroins 1 and 2 and their recombinant analogs

The distribution of secondary structure elements along the polypeptide chains of spider silk proteins spidroins 1 and 2 and their recombinant analogs has been studied by statistical methods. It was found that these proteins as monomers contain only traces of β-structure, while the Ala-rich and the Gly-rich regions are predicted as α-helices and as left-handed helices of polyproline II type. Analysis of literature and our CD data shows that the major polypeptide chain conformation of spidroins 1 and 2 and their recombinant analogs in aqueous solutions is the polyproline II helix, with some α-helices and a very small share of β-structures. The transition to the state with extended conformations, which are characteristic of mature silk fibers, requires dehydration of the polypeptide backbone. Thus, the genesis of β-structure in spider web proteins is determined by the conditions of transitions between the main regular backbone conformations.